The Transactions of the Korea Information Processing Society (정보처리학회 논문지)

Korea Information Processing Society (한국정보처리학회)
권호 표지
DOI QR코드

DOI QR Code

Development of a Deep Learning-based Midterm PM2.5 Prediction Model Adapting to Trend Changes

경향성 변화에 대응하는 딥러닝 기반 초미세먼지 중기 예측 모델 개발

  • 민동준 (고려대학교 정보보호대학원 정보보호학과) ;
  • 김혜림 (이화여자대학교 사범대학) ;
  • 이상근 (고려대학교 정보보호대학원)
  • Received : 2024.01.17
  • Accepted : 2024.05.18
  • Published : 2024.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

Fine particulate matter, especially PM2.5 with a diameter of less than 2.5 micrometers, poses significant health and economic risks. This study focuses on the Seoul region of South Korea, aiming to analyze PM2.5 data and trends from 2017 to 2022 and develop a mid-term prediction model for PM2.5 concentrations. Utilizing collected and produced air quality and weather data, reanalysis data, and numerical model prediction data, this research proposes an ensemble evaluation method capable of adapting to trend changes. The ensemble method proposed in this study demonstrated superior performance in predicting PM2.5 concentrations, outperforming existing models by an average F1 Score of approximately 42.16% in 2019, 58.92% in 2021, and 34.79% in 2022 for future 3 to 6-day predictions. The model maintains performance under changing environmental conditions, offering stable predictions and presenting a mid-term prediction model that extends beyond the capabilities of existing deep learning-based short-term PM2.5 forecasts.

초미세먼지, 특히 지름이 2.5㎛ 이하인 PM2.5는 인체 건강과 경제에 큰 피해를 주는 오염물질이다. 본 연구는 대한민국 서울 지역을 중심으로, 2017년부터 2022년까지 자료를 수집하여 PM2.5 데이터 분석 및 데이터 경향성 변화 추이를 분석하고, PM2.5 중기 예측 모델을 개발하는 것을 목표로 한다. 수집, 생산된 대기질 및 기상 데이터, 재분석 데이터, 수치모델 예측 데이터를 바탕으로, 모델을 학습하고 이를 통합한 경향성 변화에도 대응할 수 있는 앙상블 기법을 제안한다. 본 연구에서 제안하는 앙상블 기법은 PM2.5 농도 예측 성능 면에서 기존 모델 대비 미래 D+3~D+6 예측일 F1 Score 기준 평균 2019년 약 42.16%, 2021년 약 58.92%, 2022년 약 34.79% 높은 성능을 보였다. 제안한 모델은 변화하는 환경 조건에도 성능을 유지함으로써 안정적인 예측을 가능하게 하며, 기존 딥러닝 기반 PM2.5 단기 예측보다 먼 예측을 수행하는 중기 예측 모델을 제시한다.

Keywords

References

  1. WHO Health Organization, "WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide" [Internet], https://www.who.int/publications/i/item/9789240034228
  2. M. Bae, Y.-H. Kang, E. Kim, S. Kim, and S. Kim, "A multifaceted approach to explain short- and long-term PM2.5 concentration changes in Northeast Asia in the month of January during 2016-2021," Science of The Total Environment, Vol.880, pp.163309, 2023.
  3. S. Lee, et al., "Analysis of long-range transboundary transport (LRTT) effect on Korean aerosol pollution during the KORUS-AQ campaign," Atmospheric Environment, Vol.204, pp.53-67, 2019. https://doi.org/10.1016/j.atmosenv.2019.02.020
  4. Y. Cha, C.-K. Song, K. Jeon, and S.-M. Yi, "Factors affecting recent PM2.5 concentrations in China and South Korea from 2016 to 2020," Science of The Total Environment, Vol.881, pp.163524, 2023.
  5. Ministry of Environment, Artificial intelligence prediction improves ozone forecast accuracy [Internet], https://www.me.go.kr/home/web/board/read.do?menuId=10525&boardMasterId=1&boardCategoryId=39&boardId=1372380
  6. Ministry of Environment, "Expansion of Weekly Forecast Regions and Strengthening Response to the 2nd Fine Dust Seasonal Management," Air Quality Integrated Forecast Center [Internet], www.me.go.kr
  7. K. P. Singh, S. Gupta, A. Kumar, and S. P. Shukla, "Linear and nonlinear modeling approaches for urban air quality prediction," Science of The Total Environment, Vol.426, pp.244-255, 2012. https://doi.org/10.1016/j.scitotenv.2012.03.076
  8. J. Zhao, F. Deng, Y. Cai, and J. Chen, "Long short-term memory - Fully connected (LSTM-FC) neural network for PM2.5 concentration prediction," Chemosphere, Vol.220, pp.486-492, 2019. https://doi.org/10.1016/j.chemosphere.2018.12.128
  9. H. S. Kim, K. M. Han, J. Yu, J. Kim, K. Kim, and H. Kim, "Development of a CNN+LSTM Hybrid Neural Network for Daily PM2.5 Prediction," Atmosphere, Vol.13, No.12, Art. no.12, 2022.
  10. K. Lee et al., "Development of Korean Air Quality Prediction System version 1 (KAQPS v1) with focuses on practical issues," Geoscientific Model Development, Vol. 13, No.3, pp.1055-1073, 2020. https://doi.org/10.5194/gmd-13-1055-2020
  11. U. Shin et al., "Predictability of PM2.5 in Seoul based on atmospheric blocking forecasts using the NCEP global forecast system," Atmospheric Environment, Vol.246, pp.118141, 2021.
  12. A. Sayeed, Y. Lops, Y. Choi, J. Jung, and A. K. Salman, "Bias correcting and extending the PM forecast by CMAQ up to 7 days using deep convolutional neural networks," Atmospheric Environment, Vol.253, pp.118376, 2021.
  13. A. Azid et al., "Prediction of the Level of Air Pollution Using Principal Component Analysis and Artificial Neural Network Techniques: a Case Study in Malaysia," Water Air Soil Pollut, Vol.225, No.8, pp.2063, 2014.
  14. C.-J. Huang and P.-H. Kuo, "A Deep CNN-LSTM Model for Particulate Matter (PM2.5) Forecasting in Smart Cities," Sensors, Vol.18, No.7, Art. No.7, 2018.
  15. H. Chang-Hoi et al., "Development of a PM2.5 prediction model using a recurrent neural network algorithm for the Seoul metropolitan area, Republic of Korea," Atmospheric Environment, Vol.245, pp.118021, 2021.
  16. R. B. Cleveland, W. S. Cleveland, J. E. McRae, and I. Terpenning, "STL: A seasonal-trend decomposition," Journal of Official Statistics, Vol.6, No.1, pp.3-73, 1990.
  17. R. Bellman, "Dynamic programming," Science, Vol.153, No.3731, pp.34-37, 1966. https://doi.org/10.1126/science.153.3731.34
  18. S. W. Choi and B. H. S. Kim, "Applying PCA to Deep Learning Forecasting Models for Predicting PM2.5," Sustainability, Vol.13, No.7, Art. No.7, 2021.
  19. I. J. Goodfellow, et al., "Generative Adversarial Networks," Neural Information Processing Systems, 27, 2014.
  20. J. Yoon et al., "Time-series generative adversarial networks," Neural Information Processing Systems, 32, 2019.
  21. C. Szegedy et al., "Going deeper with convolutions," Computer Vision and Pattern Recognition, pp.1-9, 2015.
  22. K. Cho et al., "Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation," Empirical Methods in Natural Language Processing, pp.1724-1734, 2014.